This study focuses on development and implementation of genomic technologies for the study of cancer. Older technologies are largely inadequate to deal with complex changes in the structure of the genome and alterations of gene expression occurring during oncogenesis. Improvements in technology for placing molecular probes onto the cytogenetic, genetic and physical maps now enable the large-scale characterization of the structure and function of the disturbed cancer genome with the following approaches. 1) High-resolution positional reagents and visualization methods encompass microdissection technology as well as high-resolution fluorescence and multicolor in situ hybridization (FISH). These methods are being applied to the analysis of previously intractable problems in cancer cytogenetics. 2) A new technology, the use of cDNA microarrays, has been developed to allow simultaneous evaluation of cellular mRNA levels for thousands of genes. This technology specifically enables sensitive comparisons of gene transcript levels between cells from various pathological stages. Experiments to date with model systems such as normal cells transduced with oncogenes and staged clinical specimens have allowed detection of many changes in gene expression associated with important biological and clinical events. This system enables the global analysis of gene expression in cancer cells for the first time. It is being applied to both clinical specimens and laboratory models of cancer development and progression and has been adapted to the determination of copy number change. 3) Evaluation of results obtained with the above technologies can be examined with the new approach of tissue microarrays developed within the Branch, allowing the simultaneous examination of copy number change and gene expression (by immunohistochemistry and in situ hybridization) simultaneously in thousands of tumor specimens. Automated devices with bioinformatic support enable the unprecedented acquisition of vast amounts of data linked to biological and clinical endpoints.